PhD in Medical Electronics

1D-CNN, Transformer and hybrid models for multi-class arrhythmia classification and heart-failure prediction from single- and multi-lead ECG, with TinyML deployment on STM32/Jetson for wearable monitoring.

Spatio-temporal deep learning and graph neural networks for early seizure prediction and cognitive-state classification from multi-channel EEG, validated on public datasets and experimental recordings.

Sensor fusion of ECG, PPG, EMG and accelerometer data using Kalman filters and deep learning for robust vital-sign estimation and activity recognition in ambulatory settings.

Cadence Virtuoso design of ultra-low-noise, high-CMRR instrumentation amplifiers and ADCs for ECG/EEG/EMG, with power and noise optimisation for implantable and wearable use.

Full-wave design and SAR analysis of implantable antennas and inductive/ resonant wireless power links using ANSYS HFSS and CST, with tissue-equivalent phantom validation.

System-level design of sensing, signal processing and stimulation circuits for closed-loop deep-brain or peripheral-nerve stimulation, combining Cadence IC design with MATLAB system modelling.

Flexible multi-modal wearable platform integrating ECG, PPG, temperature and motion sensing with on-node AI inference on STM32/ESP32 and Bluetooth/LoRa telemetry.

COMSOL Multiphysics modelling of RF and electromagnetic field interaction with biological tissues, including SAR, temperature rise and safety assessment for implants and wearable transmitters.

Deep learning architectures with attention gates and transformers for accurate segmentation of MRI/CT/ultrasound images, optimised for edge inference on NVIDIA Jetson platforms.

Real-time image enhancement, artefact reduction and automated measurement pipelines for portable ultrasound systems using MATLAB and deep learning models.

Interpretable deep learning models with SHAP/Grad-CAM visualisation for ECG, EEG and medical image diagnosis, addressing regulatory and clinical trust requirements.

Multiphysics simulation of photoacoustic wave generation and propagation in tissue using COMSOL, coupled with reconstruction algorithms in MATLAB/Python.

Design of multi-channel, low-noise neural recording ASICs with multiplexing and digitisation in Cadence, targeting implantable brain-machine interface applications.

Hardware-efficient spike detection, sorting and motor-intention decoding algorithms implemented on FPGA/STM32 for closed-loop BMI systems.

RF and inductive telemetry system design for high-channel-count neural data transmission, including link budget, modulation and SAR constraints using HFSS and system simulation.

Signal acquisition, processing and machine-learning decoding from peripheral nerve or EMG for intuitive prosthetic limb control, with real-time embedded implementation.